Skip to main content

Table 6 Computational complexity of the implemented linear and non-linear prediction models

From: Genomic prediction based on data from three layer lines using non-linear regression models

 

Linear models

Non-linear models

Ridge regression

GBLUP

 

With PCA

o(r3)

o(n3)

o(n3)

Without PCA

o(m3)

o(n3)

o(n3)

  1. For linear and non-linear models both implementations with and without PCA are considered. In the table, r, m, and n indicate the number of principal components (r), genotypes (m), and training animals (n). In the case of genomic prediction, generally r ≤ n ≪ m.